On Tue, Jun 01, 2004 at 07:08:53PM -0500, Peter Eisch wrote:
> 
> I just built a kernel with every wee bit of debug enabled that I could find.
> Should I futz with load-base?  (as I recall to myself some gory days with
> MacPPCs)
> 
> ...
> 13000 >> NetBSD/sparc64 OpenFirmware Boot, Revision 1.7
> >> (peter@zipper, Tue Jun  1 12:10:45 CDT 2004)
> bootp: no reply
> Using BOOTPARAMS protocol: ip address: 12.30.65.29, hostname: e450
> root addr=12.30.65.28 path=/export/e450/root
> main: Calling loadfile(fd, /pci@1f,4000/network@1,1/netbsd -d)
> loadfile: reading header
> elf64_exec: Booting /pci@1f,4000/network@1,1/netbsd
> 4350792@0x1000000+159456@0x1800000+4034848@0x1826ee0
> symbols @ 0xfef7a440 164+309672+188731 start=0x1000000
> chain: calling OF_chain(800000, cbb8, 1000000, fff7fa80, 18)
> Type  'go' to resume
> {0} ok go
> OF_chain: prom returned!
> Calling entry(0, 0xfff7fa80, 18, f0075a4c, f0075a4c)
> Loaded initial symtab at 0xfef7a000, strtab at 0xfefc5e90, # entries 12889
> Entered pmap_bootstrap.
> Available virtual memory:
> memlist start 0xfffff80000000000 size 7fc00000000
> memlist start 0x100000000 size 7ff00000000
> memlist start 0xffff0000 size e000
> memlist start 0x1c00000 size ee400000
> memlist start 0x1428000 size 3d8000
> memlist start 0x0 size 1000000
> memlist start 0xfff1a000 size 2000
> memlist start 0xfff0e000 size 4000
> memlist start 0xfef00000 size 7a000
> memlist start 0xf0800000 size e6f4000
> End of available virtual memory
> Trying to allocate msgbuf at e0006000, size 8000
> Fast Data Access MMU Miss
> {0} ok 

You might also want to do a ctrace at this point.

Hm.  It appears to be failing trying to allocate the message buffer.

If you look in prom_get_msgbuf() you will see code to explicitly
disable some operations on e250s.  I'd speculate that e450s, which
are very similar machines, have the same problem.

Eduardo